A pentacoordinate oxorhenium(V) metallochelate elicits antibody catalysts for phosphodiester cleavage

Full text of DOI:10.1021/ja970107v

4088
J. Am. Chem. Soc. 1997, 119, 4088-4089
A Pentacoordinate Oxorhenium(V) Metallochelate
Elicits Antibody Catalysts for Phosphodiester
Cleavage
David P. Weiner, Torsten Wiemann, Mary M. Wolfe,
Paul Wentworth, Jr., and Kim D. Janda*
Department of Chemistry, The Scripps Research
Institute and The Skaggs Institute for
Chemical Biology, 10550 North Torrey Pines Road
La Jolla, California 92037
Figure 1. Comparison of the putative transition state for RNase
A-catalyzed RNA cleavage and the oxorhenium(V) transition-state
analogues (1a, 1b).
ReceiVed January 15, 1997
Scheme 1
The cleavage of phosphodiester bonds, such as those found
in DNA and RNA, is a reaction of key importance in living
systems. Consequently there are intense efforts to develop novel
phosphodiesterases for use in biochemistry and medicine.1 Our
approach to this problem has been to exploit the diversity of
the vertebrate immune system2 to generate antibodies with
phosphodiesterase activity, focusing initially on antibody ca-
talysis of RNA cleavage.3 The most-studied natural ribonu-
clease is RNase A which catalyzes the hydrolysis of the P-O5′
bond of RNA in two steps: attack on phosphorus by the 2′-
hydroxyl group to form a 2′, 3′-cyclic phosphodiester intermedi-
ate, followed by hydrolysis of this intermediate to give a 3′-
phosphomonoester (Scheme 1).4
While there is currently a lively debate on the precise
mechanism of this enzyme,5 it is generally accepted that both
the ring closure and ring opening steps involve a similar
pentacoordinate negatively-charged transition state (TS) (Figure
1). Classically this TS has been assumed to be trigonal
bipyramidal (TBP)6 though several studies have suggested more
of a distorted TBP/square pyramidal (SP) geometry.7 The
challenge in designing mimics of this TS for use as haptens
reduces to representing its geometry and charge characteristics
in a stable molecule. Vanadates would be promising candidates
and have been shown to be potent inhibitors of several
phosphoryl transfer enzymes including RNase A.8 However,
vanadium alkoxides undergo rapid ligand exchange in aqueous
solution, making them unsuitable for immunization.9 Recently
we have shown that oxorhenium(V) (oxoRe(V)) and oxotech-
netium(V) adenosine complexes are good inhibitors of the
purine-specific ribonuclease, RNase U2.10 These metallochelates
possess a water-stable negatively-charged pentacoordinate struc-
ture, most likely existing in a distorted TBP/SP geometry11
(Figure 1). Therefore we reasoned that oxoRe(V) uridine
complexes 1a and 1b could be mimics of the TS for either the
first or second step in RNA hydrolysis and thus be promising
haptens for the generation of catalytic antibodies for this
reaction.
Hapten 1 was synthesized as a pair of diastereomers in ten
steps starting from 1-â-D-arabinofuranosyluracil.12 The anti and
syn epimers (1a and 1b) were separated by preparative reverse-
phase HPLC and coupled to keyhole limpet hemocyanin (KLH)
through the 5′-hydroxyl group of the ribose ring using glutaric
anhydride. Mice were immunized with the hapten-KLH
conjugates, and 50 hapten-specific hybridoma cell lines (25 each
for 1a and 1b) were isolated using standard hybridoma
technology.13 Monoclonal antibodies (IgG) were purified to
homogeneity from tissue culture and ascites fluid.14 All 50
antibodies were then screened for their ability to catalyze the
cleavage of uridine 3′-(p-nitrophenyl phosphate) (UpOC6H4-p-
NO2)15 (Scheme 2).
(1) (a) Chin, J. Acc. Chem. Res. 1991, 24, 145-152. (b) Magda, D.;
Miller, R. A.; Sessler, J. L.; Iverson, B. L. J. Am. Chem. Soc. 1994, 116,
7439-7440. (c) Lorsch, J. R.; Szostak, J. W. Acc. Chem. Res. 1996, 29,
103-110.
(2) Recent reviews on catalytic antibodies: (a) Janda, K. D.; Shevlin,
C. G.; Lo, C.-H. L. In ComprehensiVe Supramolecular Chemistry;
Murakami, Y., Ed.; Pergamon: Oxford, 1996; Vol. 4, pp 43-72. (b) Schultz,
P. G.; Lerner, R. A. Science 1995, 269, 1835-1842.
(9) (a) Paulsen, K.; Rehder, D.; Thoennes, D. Z. Naturforsch. 1978, 33A,
834-839. (b) Tracey, A. S.; Gresser, M. J.; Liu, S. J. Am. Chem. Soc.
1988, 110, 5869-5874. (c) Crans, D. C.; Rithner, C. D.; Theisen, L. A. J.
Am. Chem. Soc. 1990, 112, 2901-2908. (d) Crans, D. C.; Felty, R. A.;
Miller, M. M. J. Am. Chem. Soc. 1991, 132, 265-269.
(3) Wentworth Jr., P., Janda, K. D. SYNLETT In press.
(4) (a) Markham, R.; Smith, J. D. Biochem. J. 1952, 52, 552-557. (b)
Brown, D. M.; Todd, A. R. J. Chem. Soc. (London), 1953, 52-58. (c)
Thompson, J. E.; Venegas, F. D.; Raines, R. T. Biochemistry 1994, 33,
7408-7414.
(10) For previous work from our laboratory on the inhibition of
ribonuclease U2 by oxorhenium and oxotechnetium complexes, see: (a)
Chen, Y.-C. J.; Hansske, F.; Janda, K. D.; Robbins, M. J. J. Org. Chem.
1991, 56, 3410-3413. (b) Chen, Y.-C. J.; Janda, K. D. J. Am. Chem. Soc.
1992, 114, 1488-1489. (c) Wentworth Jr., P.; Wiemann, T.; Janda, K. D.
J. Am. Chem. Soc. 1996, 118, 12521-12527.
(5) (a) Breslow, R.; Xu, R. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 1201-
1207. (b) Breslow, R.; Dong, S. D.; Webb, Y.; Xu, R. J. Am. Chem. Soc.
1996, 118, 6588-6600. (c) Breslow, R.; Chapman, Jr., W. H. Proc. Natl.
Acad. Sci. U.S.A. 1996, 93, 10018-10021. (d) Thompson, J. E.; Raines, R.
T. J. Am. Chem. Soc. 1994, 116, 5467-5468. (e) Herschlag, D. J. Am.
Chem. Soc. 1994, 116, 11632-11635.
(11) For leading references on other oxorhenium complexes, see: (a)
O’Niel, J. P.; Wilson, S. R.; Katzenellenbogen, J. A. Inorg. Chem. 1994,
33, 319-323. (b) Yoon, C. D.; Wilson, S. R.; Katzenellenbogen, J. A. Inorg.
Chem. 1995, 34, 1624-1625.
(6) (a) Usher, D. A.; Richardson, D. I.; Eckstein, F. Nature (London)
1970, 228, 663-665. (b) Usher, D. A.; Erenrich, E. S.; Eckstein, F. Proc.
Natl. Acad. Sci. U.S.A. 1972, 69, 115-118.
(7) (a) Thatcher, G. R. J.; Kluger, R. AdV. Phys. Org. Chem. 1989, 25,
99-265. (b) Holmes, R. R.; Dieters, J. A.; Galluci, J. C. J. Am. Chem. Soc.
1978, 100, 7393-7402. (c) Lim, C.; Tole, P. J. Am. Chem. Soc. 1992, 114,
7245-7252.
(12) Overall yield for the hapten synthesis was 9.4%; details of the
synthesis will be published elsewhere. All new compounds exhibited
1
satisfactory 13C and H NMR and mass spectral analysis.
(13) Kohler, G.; Milstein, C. Nature (London) 1975, 256, 495-497.
(14) Antibodies were purified by ammonium sulfate precipitation, DEAE
ion exchange, and protein G affinity chromatography. At this stage
antibodies were >99% pure based on silver staining of SDS-PAGE gels.
Monoclonal cell lines are available on request.
(8) (a) Crans, D. C.; Bunch, R. L.; Theisen, L. T. J. Am. Chem. Soc.
1989, 111, 7597-7607. (b) Linquist, R. N.; Lynn, J. L.; Lienhard, G. E. J.
Am. Chem. Soc. 1973, 95, 8762-8768. (c) Leon-Lai, C. H.; Gresser, M. J.;
Tracey, A. S. Can. J. Chem. 1996, 74, 38-48.
S0002-7863(97)00107-8 CCC: $14.00 © 1997 American Chemical Society

Communications to the Editor
J. Am. Chem. Soc., Vol. 119, No. 17, 1997 4089
Scheme 2
position of the sugar, were also investigated as inhibitors20 and
found to behave in a similar way to 1a and 1b. This suggests
that the antibody primarily recognizes the Re core common to
both classes of compounds but does not markedly distinguish
between the two configurations of the Re oxygen. The apparent
lack of importance of the base for inhibition in the antibody
10c
contrasts with experiments on RNase U2 where oxoRe(V)
complexes of 9-(2′,3′-diamino-2′,3′-dideoxy-â-D-ribofuranosyl)-
adenine were found to be much more potent inhibitors of the
enzyme than 2a and 2b. However, between anti and syn
diastereomers there was only a small (6-fold) difference in Ki.10c
The lack of strong diastereoselectivity in the antibody and
enzyme may reflect the fact that the negative charge of the Re
complex is delocalized over the diamide dimercaptide ligand
system reducing the impact of the Re oxygen configuration in
molecular recognition.
Antigen-binding fragments (Fab) of 2G12 were prepared and
showed indistinguishable kinetic properties to the parent
antibody, while the constant region fragments (Fc) exhibited
no catalysis. This further supports the notion that catalysis takes
place in the antibody binding site.
We have performed an experiment analogous to immuno-
precipitation21 but using immobilized protein G as the precipi-
tant. By treating identical solutions of IgG 2G12 with varying
amounts of immobilized protein G and filtering the samples
through 0.22 µm spin-filters, it was possible to selectively
remove varying amounts of antibody from the solutions.
Analysis of the protein G-treated filtrates showed that catalytic
activity was linearly proportional to the remaining protein
concentration providing perhaps the most direct evidence that
catalysis is due to the antibody. This is a relatively simple
experiment to perform and offers a general and powerful control
for all antibody-catalyzed reactions where the possibility of a
contaminating enzyme exists.
In this communication we have reported the isolation of
antibodies that cleave a phosphodiester bond. This finding
further supports our hypothesis that oxoRe(V) complexes can
be effective transition-state analogues for phosphoryl transfer
reactions. Work is underway to isolate more efficient phos-
phodiesterase antibodies by improving on the original hapten
design. Two ways that this might be achieved are by minimiz-
ing the size of the Re ligand system and/or by incorporating a
leaving group mimic into the ligand design.3
Three antibodies (all raised to the anti hapten, 1a) were found
to catalyze the reaction. These antibodies all lost activity upon
boiling for 3 min, mitigating against contamination with
thermostable RNases. Moreover, their activities were unaffected
by the addition of 1 mM EDTA excluding the possibility of
contamination by a metal-dependent phosphodiesterase.
The most active antibody (2G12) was studied in more detail.
The greatest catalytic rate enhancement by this antibody was
observed at pH 6.0, yielding a kcat of 1.53 ( 0.09 × 10-3 s-1
and a KM of 240 ( 30 µM. The reaction was typically followed
for several turnovers with no indication of product inhibition.
Base-catalyzed cleavage of UpOC6H4-p-NO2 was found to be
first-order with respect to hydroxide ion. A small linear
dependence on buffer concentration was measured, while the
antibody-catalyzed reaction was independent of buffer effects.16
Extrapolating to zero buffer concentration yielded kuncat ) 4.9
× 10-6 s-1 and kcat/kuncat ) 312.17 The catalytic proficiency of
an enzyme can be expressed as [(kcat/KM)/kuncat]
18 which for Ab
2G12 is 1.3 × 106 M-1. This compares to a value of 1.1 ×
1010 M-1 for RNase A at pH 6.0 using the same substrate.5d
When studying any artificial catalyst where natural enzymes
catalyze the same reaction, it is essential to exclude the
possibility of enzyme contamination. This is especially impor-
tant for the cleavage of phosphodiesters. Several experiments
have been performed that together unequivocally rule out
enzyme contamination, the most important of which are hapten
inhibition, retention of specific activity by Fab fragments, and
protein G immunoprecipitation, Vide infra.
The hapten (1a) was found to be a tight-binding inhibitor of
Ab 2G12. An IC50 value of 6 µM was measured by varying
the inhibitor concentration in reactions containing 5 µM antibody
(10 µM active sites) and 400 µM substrate. Assuming a
competitive mode of inhibition this gives an upper limit to the
Ki of 0.4 µM.19 Interestingly both the anti (1a) and syn (1b)
epimers of the hapten were found to give indistinguishable IC50
values. The oxoRe(V) complexes of 2,5-anhydro-3,4-diamino-
3,4-dideoxy-D-ribitol (2a and 2b), lacking a base at the 1′
Acknowledgment. This work was supported by the National
Institute of Health and by the Skaggs Institute for Chemical Biology.
T.W. thanks the Alexander von Humboldt foundation (Feodor Lynen
program) for support.
(15) Uridine 3′-(p-nitrophenyl phosphate) was synthesized according to
Davis, A. M.; Hall, A. D.; Williams, A. J. Am. Chem. Soc. 1988, 110,
5105-5108, but using a triethylammonium bicarbonate wash in the workup
to give the triethylammonium salt, see: Vandersteen, A. M. Ph.D. Thesis,
Kings College, London, 1995; pp 355-356.
(16) Reaction conditions were 10 µM Ab active sites, 20 mM 4-mor-
pholineethanesulfonic acid (MES) pH 6.0. The second-order rate constant
for buffer catalysis is 2.04 × 10-4 s-1 M-1. Thus the ratio of the kcat/KM
for Ab 2G12 to the second-order rate constant for buffer catalysis is 3.1 ×
104.
Supporting Information Available: Buffer effect on the cyclization
of UpOC6H4-p-NO2, Michaelis-Menten analysis of Ab 2G12, details
of the protein G immunoprecipitation, and procedures for antibody
preparation and purification (5 pages). See any current masthead page
for ordering and Internet access instructions.
(17) Our value of kuncat is consistent with recently published kinetics:
Dantzman, C. L.; Kiessling, L. L. J. Am. Chem. Soc. 1996, 118, 11715-
11718.
JA970107V
(18) Radzicka, A.; Wolfenden, R. Science 1995, 267, 90-93.
(19) Copeland, R. A.; Lombardo, D.; Giannaras, J.; Decicco, C. P. Bioorg.
Med. Chem. Lett. 1995, 5, 1947-1952.
(20) For the synthesis of 2a and 2b see ref 10c.
(21) Harlow, E.; Lane, D. Antibodies: a laboratory manual; Cold Spring
Harbor Laboratory: New York, 1988.